A First Look at MLS Teams' 2025 Performance
What was that? Something about the sample size? Couldn't hear you, I'm sure it's fine.
In the preseason I wrote a series of articles categorizing MLS teams based on their 2024 performance (part 1, part 2, and part 3) and I’ve been using those in preview articles like this one from before week 4.
As we get farther into the season, 2024 is becoming less and less relevant with all the player and coaching changes. Four 2025 games isn’t a huge sample size, but it’s enough to give us something to go on, so let’s see if we can decide how to regroup teams. Along the way, maybe we can learn a thing or two about how much MLS teams change from year to year.
Strength of Schedule
The massively uneven opponent quality is the biggest problem with trying to interpret these numbers so we’ll just confront that right away. I measured this by considering how many points a team’s opponents have earned not including the given team. This means the sample is even smaller—three games—than the other 2025 articles. But let’s take a look:
Some good news here for struggling teams like the LA Galaxy and the New England Revolution. It’s been tough sledding so far but it will probably get easier. San Diego, on the other hand, can feel proud of their good start.
Meanwhile, for Western teams looking up at Vancouver’s maximum points, there’s some hope here that they’ll revert to something a bit less dominating in time.
DC United is undefeated, but they only have one win and have looked consistently unconvincing. Their next few games will be big test of MLS parity. How much does it matter they’ve played by far the easiest schedule by this measure? You’d think a lot…but this entire newsletter is based on struggling to discern differences in MLS teams. How big is this difference in practice?
Offense
For the season-long 2024 offense assessments, I used actual goals scored to rate offenses instead of expected goals. I still think that was the right choice. Even across a season’s worth of games, there are big discrepancies between actual goals scored and expected goals, and there’s reason to think that some teams (e.g. Miami) dependably differ from expectations.
However, since we’re looking at just 4 games here, actual goals involve a huge amount of luck. A single own goal might be the difference between a “good” vs. “great” offense, and then there’s penalties.
So at least for now, I’m going to use expected goals. Specifically, non-penalty expected goals. Getting a penalty awarded is about as lucky as a low-percentage shot going in, I think, so the offenses that are producing the most npxG are the ones generating the most dangerous shots.
I know this isn’t perfect. Four games worth of npxG will still be heavily distorted from both varying opponent quality and by good or bad luck, like playing a lot of minutes against a team down a man or having a star player miss a game due to a minor ailment.
But there’s no perfect measure. So what do we end up with?
If we were to do the same thing with actual goals, the teams would be rearranged but there wouldn’t be big changes. Philadelphia, Miami, Chicago, and Orlando are the only teams out-performing their xG by at least two goals and they are close to the top of the list. Chicago and Miami would move into the top category, certainly but Philly and Orlando are already there. Penalty xG isn’t counted here but no team has scored more than one penalty, though DC United has earned two and Nashville earned three.
On the other side, there’s three teams underperforming xG by at least two goals: Austin (-2.2), Red Bulls (-2.6), and Montréal (-2.7). They’d probably be in lower categories if this were actual goals.
Now let’s compare 2024 to 2025 so far:
Always remembering the caveat that it’s just four games, I think there’s a few interesting conclusions to draw.
First, one key conclusion of preseason punditry is validated. Is it bad to lose offensive talent and not replace them? Columbus Crew, LAFC, and LA Galaxy lost talent (in large part to injury in the Galaxy’s case, though they also lost Joveljić for good) and sure enough their offenses are worse. NYCFC has had a lesser dropoff after losing, with apologies to Santi Rodríguez, something of a lesser talent.
Then there’s the real core precept of preseason punditry: adding talent in the offseason means getting immediately better. Cincinnati and Portland were congratulated for immediately replacing Acosta and Evander, Charlotte praised for bringing in Zaha, and of course everyone was sure Atlanta’s big spending ways would be rewarded. These moves might still pay off as the big name players and the teams around them adjust, but at least we can say the impact has not been instant.
The two teams that do seem to have gotten much better on offense, San Jose and Chicago, did so by firing everyone and bringing in a “big MLS name” to be both coach and general manager. The other team with a similarly ambitious rebuild project, New England, did this with, hmm, some medium MLS names and isn’t having anything like the same success.
But one reason to expect changes is how bottom-heavy this graph is. The average team is worse on offense so far. I can’t explain it definitively, but I have a few plausible theories:
Team cohesion takes time, especially with lots of big changes this offseason. Some of these teams will generate more opportunities as their players gel with each other and the manager.
As the season wears on, injuries mount and the temperature rises, slowing down games. To be honest, I would have guessed both of these things would favor the defense, but maybe defenses get thin and tired and it actually helps offenses?
In some absolute sense, the talent got worse. Maybe Cucho and Thiago Martinez (who contributes a fair number of games in 2024 for Atlanta) were just better attacking players than David da Costa and this version of Almirón? I don’t really believe this one, but I’m listing it to be comprehensive. I suppose if it’s true, we won’t see much improvement here until the midseason transfer window.
Probably a lot could be learned about this by looking at past seasons, but my time to work on this newsletter is far from infinite, so hey, it’ll be very interesting to keep an eye on this as we go through 2025.
Defense
Now let’s turn to defense. Again, I think it makes sense to use expected goals rather than actual goals given how few games we’re talking about. I think there’s a better case here for including penalties, since it feels like good luck for the offense when an attacker gets tripped in the corner of the box but a bad play by the defense. However, I think I’ll stick with non-penalty expected goals since it seems like it might be more predictive. Chicago and Portland, the only two teams to give up two penalties, can breathe a sigh of relief, but they need to stop conceding penalties now or they’re going to make me look bad.
So here’s how things look:
Orlando and Toronto have each given up 10 goals, two more than any other team. Using npxG here lets Orlando slip into the merely “Bad” category, but Toronto looks much different. Their opponents have outshot their xG by a staggering 5.0 goals, leading the league in that stat and almost doubling Dallas and Houston in second place (2.6). I thought Toronto looked better than I expected in the game I saw them play, so maybe they have just been unlucky?
On the other side of things, St. Louis has drawn lots of attention for somehow starting the season with four clean sheets, but they didn’t even make my top category. Vancouver and Austin have each given up 2 goals and this suggests that could be more sustainable, whereas San Diego, also sitting on 2 goals allowed, has been a little lucky.
Now let’s look at the changes:
For the most part this is the kind of chart I think you’d expect. Big improvements from St. Louis and Austin and a major regression from Real Salt Lake. My takeaway is that for the most part there just aren’t any clear correlations. Some teams got better, some got worse, and while I’m sure there are reasons for these changes, I don’t see sweeping general explanations.
Dallas, LAFC, and LA Galaxy were in my “high” category for transfer spending. Hasn’t helped so far. Kansas City and Toronto didn’t spend anything significant but improved (in xG terms anyway). The rebuild operations at San Jose and Chicago I was praising in the Offense section don’t seem to have accomplished much, whereas New England is looking much better on defense so far.
Shotstopping
I won’t explain the shotstopping stat again (you can see my first preseason article for that), so I’ll just note that unlike the previous sections, this includes penalty kicks.
Probably a lot of misleading random variation in these stats. I don’t have the time to attempt to prove it, but I think PSxG is more noisy than more common expected stats like xG and xA. If nothing else, the early returns look good for San Diego, a team that went very cheap at the goalkeeper position during their build-out. I was skeptical about this but so far it’s working.
Let’s look at the 2024/2025 delta in PSxG - GA:
San Jose has the best improvement with their starter, Daniel, back after spending much of 2024 injured. There were no changes in Colorado, just Zach Steffen having a huge turnaround season—so far! DC United was the only team to invest a small transfer fee into the position by signing U22 goalkeeper Kim Joon Hong and that is paying off so far.
Houston’s decision to move on from Steve Clark is really not working yet. Sporting Kansas City and LA Galaxy have likewise not seen the improvement they might have hoped for from a new starter.
It might be time for some concerns about Toronto and LAFC’s old goalkeepers, but the oldest goalkeeper, Brad Guzan, is still keeping pace with his previous season’s pretty good form for Atlanta.
Coming Soon
There’s more stats I’d like to do a first 2025 look on: DP performance, other attacking players, and team playing styles for example. I don’t have time this week, but I’ll hopefully be back soon with a look, so make sure to subscribe to get that when it’s ready. Meanwhile, this information will be incorporated into my week 5 previews.